A SPECTROSCOPIC SURVEY OF FAINT COMPACT OBJECTS

We present results from a low-resolution (73 angstrom) spectroscopic survey of faint compact objects conducted with the Low Dispersion Survey Spectrograph at the Anglo-Australian Telescope. The sample is a random subset of all compact sources with 22.5 less-than-or-equal-to b(J) less-than-or-equal-to 23.5 in three high-latitude fields. Additional data on compact sources with 21 less-than-or-equal-to b(J) less-than-or-equal-to 22.5 have been selected from the higher resolution (13 angstrom) survey previously published by Colless et al. (1990), and a subset of these brighter objects has been studied at both dispersions. The new fainter survey was motivated by the possibility of discovering rare, but potentially important, populations of hitherto unidentified compact sources and establishing limits on their contributions to the source counts. Low-dispersion spectra have been secured for 117 new sources, of which spectroscopic identifications are possible for 67. Incompleteness arises primarily because only certain types of object have spectral features recognizable at this low dispersion. The bulk of the identified sources are intermediate- and late-type stars but a few galaxies and QSOs are also found. Completeness is a strong function of optical colour however, permitting us to derive important new constraints on the abundance of three populations at faint magnitudes as follows. (1) We demonstrate that our survey is sensitive to the presence of typical QSOs with broad emission features, and obtain limits on the surface density of QSOs at fainter magnitudes than previously possible. Our estimates support a flattening of the count-magnitude relation beyond b(J) = 20, as indicated by earlier work at brighter magnitudes. This result is especially significant since our sample is selected only by apparent magnitude, and is thus more inclusive than previous studies which have relied on additional selection criteria such as colour or variability. (2) For faint Galactic stars, we find a colour-magnitude distribution that is consistent with models invoking an intermediate population II thick disc. More complete and extensive data would provide numerical constraints on such models. (3) Finally, we investigate whether locally rare populations may make significant contributions to the very faint (B approximately 27) source counts. The paucity of very blue compact sources in the LDSS survey suggests the abundance of such objects at fainter limits arises via a gradual increase in the star formation rates in the galaxy population, rather than via a Euclidean increase due to a compact local population or a sudden onset with apparent magnitude of very distant sources. The absence of red objects in the very faint colour distribution, together with the dominant population of Galactic stars in the LDSS sample, limits any significant contribution from dwarf stellar systems such as intergalactic globular clusters and dE galaxies.